M
I am attempting to evaluate the effect of two different environments on test results (ambient vs controlled test lab environment). We want to determine if there will be a change is sample average and standard deviation. We make polyurethane foam seating product so the environment can effect the curing process and also the measurement repeatability of hardness results.
When developing the test plan for this study, how important is sample selection in performing the study? A couple of options we have is
To me this is a mini DOE, where you need to minimize outside variation so the variation in the study results are isolated, as much as possible, to the two variables being evaluated. My concern is that using actual products introduces a certain amount of unknown variability that can negatively effect the study results. There are components that molded with the parts that could have its own independent effort on end results. To me, the sample parts going into the study needs to be as similiar as possible. Again the intent of the study is to evaluate the effect of two different variables (ambient vs test lab)
People on the other side of the debate doesn't feel that similiar physical properties are important to the study.
If I am wrong please correct me and offer how you would develop the test plan.
If I am right please describe how I can best convince folks on proper study preparation.
When developing the test plan for this study, how important is sample selection in performing the study? A couple of options we have is
- Pull 10 consecutive actual parts without concern on hardness variability within the sample. 5 parts would condition inside the controlled lab environment. The other 5 pieces would condition in the ambient envirnment.
- Pull samples using test blocks trying to assure that physical properties are as similiar as possible. 5 parts would condition inside the controlled lab environment. The other 5 pieces would condition in the ambient envirnment.
To me this is a mini DOE, where you need to minimize outside variation so the variation in the study results are isolated, as much as possible, to the two variables being evaluated. My concern is that using actual products introduces a certain amount of unknown variability that can negatively effect the study results. There are components that molded with the parts that could have its own independent effort on end results. To me, the sample parts going into the study needs to be as similiar as possible. Again the intent of the study is to evaluate the effect of two different variables (ambient vs test lab)
People on the other side of the debate doesn't feel that similiar physical properties are important to the study.
If I am wrong please correct me and offer how you would develop the test plan.
If I am right please describe how I can best convince folks on proper study preparation.